The hypothesis that brainspace constrains learning potential is not easily testable in most systems due to the lack of dedicated neural network for a single behavior. In songbirds, such a dedicated series of nuclei exist, thus facilitating direct investigation of brain/behavior relationships. Both across species and among individuals of several species, repertoire size is positively correlated with the size of song nucleus, the HVc. In zebra finches not only HVc volume, but also the number of HVc neurons predicts how much song material is accurately reproduced. We propose a three-tier approach to better understand the relationship between network space and learning potential. First, retrograde tracing will be used to identify the neuronal subpopulations within the HVc responsible for the correlation to learning. Because only a subset of these cells are added during song learning, this information will provide insight into which stages of learning may be limited by HVc neuron number, as well as further define the role of the different pathways in song learning. Second, we will begin to explore variables that might generate the tremendous variation in HVc neuron number by determining if it covaries with clutch order and/or yolk concentrations of testosterone. Finally, we will use microinfusions of the neurotrophin bFGF to increase the number of HVc neurons, testing the causality of the observed relationship. This work has broad implications for understanding the biological processes that regulate individual learning capacity.